This article is based on the speech delivered by Att. Bengü Coşkun at the Legal Generative AI Summit held in İstanbul on December 3, 2024, and is prepared with the support of AI.
Artificial intelligence is driving significant changes across industries, economies, and legal systems worldwide. Its expanding influence, as any groundbreaking technology, necessitates a closer examination of how to regulate its development and application to ensure accountability, ethical use, and market stability.
As AI technologies become more sophisticated, policymakers face the challenge of crafting regulations that strike a balance between fostering innovation and safeguarding fundamental rights. Different jurisdictions have taken varying approaches to address the regulatory needs of AI systems. In this global effort to incorporate policies which integrate AI to the legal system, Türkiye’s current regulatory landscape, recent developments, and future expectations reveal a dynamic and evolving framework that reflects global trends while incorporating local considerations.
1. Jurisdictional Approaches to AI Regulation
Globally, there are two predominant approaches to AI regulation: the centralized approach and the decentralized approach.
1.1 Centralized Approach
The centralized approach is defined by the adoption of a standalone, comprehensive AI act that establishes a distinct regulatory framework and designates sector specific supervisory authorities to ensure compliance. This framework is notably casuistic and thorough, as exemplified by the EU AI Act. The EU AI Act contains high number of detailed provisions, each addressing specific aspects of AI development, deployment, and usage. The primary objective is to ensure strong oversight, effective risk management, and accountability, particularly for high-risk AI systems.
1.2 Decentralized Approach
The decentralized approach, on the other hand, does not rely on a single, all-encompassing AI act. Instead, AI-related provisions are integrated into various sector-specific laws, such as data protection and product liability regulations. This approach allows for more flexibility and adaptability but often lacks the cohesiveness of a centralized system. Under this model, no single regulatory body oversees AI compliance. Instead, oversight is spread across multiple authorities with sector-specific expertise.
Examples of jurisdictions following the centralized approach include the EU, China and Canada, while several other countries, such as Switzerland, UK and Australia have leaned toward a decentralized approach, addressing AI-related issues within existing sectoral regulations.
2. Türkiye’s Approach to AI Regulation
Türkiye has thus far adopted a decentralized approach to AI regulation. Rather than introducing a standalone AI Act, the country has incorporated AI-related provisions into specific pieces of legislation. These provisions focus primarily on safety measures, data protection, and the safeguarding of fundamental human rights.
However, as detailed below, a change may be on the horizon. A proposal for a standalone AI Act (“AI Act Proposal”), which is currently under review by parliamentary commissions, was submitted to the Grand National Assembly of Türkiye in June 2024.
3. Key Developments in Türkiye’s AI Regulatory Framework
3.1 National AI Strategy and 2024-2025 Action Plan
Türkiye’s National AI Strategy for 2021-2025 outlines the country’s vision for AI development and measures for its implementation. This strategy is supplemented by a 2024-2025 Action Plan, which sets out 71 specific actions to be undertaken to support AI development and regulation. Notably, the enactment of a standalone AI Act is not included in these 71 actions, reflecting a preference for the decentralized approach.
To further strengthen its AI regulatory framework, the Turkish Parliament established an AI Research Commission in October 2024. Although the Commission has yet to appoint its members to conduct the research and subsequently issue a report, the findings of this research are expected to affect Türkiye’s future regulatory strategy.
3.2 The AI Act Proposal
The AI Act Proposal, which was submitted to the Parliament in 24th of June, was referred to the relevant parliamentary commissions the day after its submission but despite procedural requirements for the commissions to prepare a report within 45 days, no report has been issued yet.
The AI Act Proposal introduces core principles, and a risk-based approach to ensure security, transparency, and accountability in AI operations. It requires high-risk AI operators to adopt special precautions and register with authorities. However, the proposal faces criticism for its vague risk assessment process, unclear supervisory powers, and disjointed sanctions, which seem to mirror the EU AI Act without local adaptation.
Although the current version of the proposal is deemed unlikely to be adopted due to its failure to address concrete and specific needs, as well as an unfavorable political landscape, it is still considered a significant step as it has paved the way for parliamentary-level discussion on the issue.
4. Compliance with the EU AI Act for Türkiye’s Market Actors
As industry stakeholders watch closely as Türkiye’s AI regulatory framework continues to evolve, it is just as important for local market actors to be mindful of the extraterritorial application of the EU AI Act. If the output of an AI system is used within the EU, the EU’s AI Act may apply to Türkiye-based providers, regardless of local regulatory requirements. This is particularly relevant for Türkiye’s exporters of AI products and services.
The EU’s AI Act includes a two-year transitional period starting from August 2024, with varying obligations taking effect at different times. Local market actors should proactively assess whether their AI systems’ outputs are used in the EU, identify applicable compliance obligations, and prepare for these requirements in advance.
5. Conclusion
Türkiye’s journey toward AI regulation is still unfolding. The decentralized approach remains dominant, though proposals for a centralized AI act have sparked important discussions. The establishment of the AI Research Commission suggests that further developments are on the horizon.
Local market actors should remain vigilant regarding both domestic and EU regulatory developments. Compliance with the EU’s AI Act could become a necessity for businesses engaged in cross-border operations.
Türkiye’s legislative developments, such as the AI Research Commission’s findings, will be pivotal in shaping the future of AI regulation. For now, continued adaptation and preparedness remain essential for all stakeholders navigating the evolving AI landscape in Türkiye.
